DeepMind podcast: Zoubin Ghahramani on why uncertainty math may be AI's missing piece

ZoubinGhahrama1 · x · 2026-08-27

A new episode of the official Google DeepMind podcast features Zoubin Ghahramani, Cambridge professor and VP of Research at Google DeepMind, on AI and uncertainty.

Ghahramani has spent 30 years pioneering intelligence built on the mathematics of uncertainty. He argues that today's AI gives absolute answers with unwavering confidence even when wrong — missing a fundamental human trait: self-doubt. The episode covers the role of uncertainty, correctness vs confidence, Bayesian thinking in AI, real-world applications, and what uncertainty research means for AGI.

The 45-minute episode is available on Apple Podcasts, Spotify, and YouTube.

Related event: DeepMind Podcast: Ghahramani on the Mathematics of Uncertainty in AI(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →